Group IV semiconductors, such as silicon, germanium, and their alloys, have indirect energy bandgaps, which inhibit band-to-band radiative recombination of electrons and holes. As a result, the predominant recombination mechanism for holes and electrons in these types of semiconductors typically is non-radiative recombination at, for example, defect sites in bulk regions and at surfaces. For this reason, most devices that are formed of these types of semiconductors are inefficient emitters of light.
Silicon is the semiconductor of choice for fabricating electronic devices. It is inexpensive to work with and has a native oxide that provides superior performance and readily may be incorporated in electronic devices wherever it is needed. The significant advantages of using silicon to fabricate electronic devices have spurred many efforts to integrate light emitting devices with silicon electronics.
One approach for integrating light emitting devices with silicon involves forming light-emitting devices from direct bandgap compound semiconductors that are grown on silicon substrates. This approach, however, is not compatible with most silicon device fabrication processes (e.g., CMOS fabrication processes) due to the different thermal requirements of these processes and the processes that are used to fabricate compound semiconductor devices. This approach also may be cost-prohibitive.
Other approaches for integrating light emitting devices with silicon have focused on improving the emission efficiency of silicon. Among the approaches that have shown some promise in this regard are: use of silicon nanostructures, such as porous silicon, to form electroluminescent devices; use of silicon doped with rare-earth metals, such as erbium and cerium, which exhibit luminescent transitions in silicon and porous silicon; use of dislocations that increase the silicon bandgap by introducing local tensile strain to prevent electrons from reaching non-radiative defect sites; and the incorporation of erbium-doped silicon nanocrystals into a silicon dioxide matrix to achieve radiative recombination of carriers with reduced problems from thermal quenching, which is typical of erbium-doping in bulk silicon.
The above-described approaches have demonstrated some success in generating light from silicon devices. The light-emission efficiencies achieved by these approaches, however, are typically insufficient to displace compound semiconductors in optoelectronics applications.